Without question, speech
technologies will change this decade for the better. Speech is the first
mode of communication we learn when we are children and it remains the most
natural communication interface for humans. I'm not saying that speech
technologies will make keyboards disappear anytime soon, but there are a
host of applications that lend themselves to speech, and this is where the
excitement lies for the industry.
The first area in which speech is a necessity is IVR enhancement. My wife
couldn't find her American Express card today. I called American Express and
tried to get a live operator on the line as soon as possible, as my mind was
busy running through lurid scenarios in which the card, in criminal hands,
was being taken on a spending spree up and down Fifth Avenue. In a frantic
state, I attempted to reach the correct department. I had to hang up and
start again three times during the process because I kept pressing the wrong
buttons. I was angry and frustrated by the time a human answered. This is a
prime example of a process that cries out for a next-gen speech recognition
application.
Speech technologies have the potential to change the way we interact with
businesses and other organizations. Agents are on the front lines of our
companies, yet we pay them just over minimum wage and expect them to present
the polished image of a Fortune 1000 company. (Please don't send me nasty
letters, I'm not implying that minimum wage workers can't project the right
image. Most of them certainly can. It is the ones that can't who destroy
customer relationships forever.)
Speech technologies allow us to be notified if a conversation contains
profanity, competitors' names or even voice stress that is higher in decibel
than a certain preset threshold. These are the conversations that every
company must know about so it can rectify the damage done immediately or,
better yet, escalate the calls to a supervisor before they get totally out
of control. Speech will change the way contact centers operate over the next
decade.
If you think that Indian contact centers are a threat to American call
center jobs, just wait until speech comes fully into its birthright. At
first, speech will take away some of the repetitive, boring jobs ('What's my
balance?'), allowing agents to migrate up the food chain and take more
interesting calls. Many companies will allow customers talking to speech
systems to immediately connect to a live agent for the purpose of answering
more complex questions or even for cross-selling and upselling purposes.
Other companies will just reduce headcount.
Still, we need to look at speech as an opportunity in the same manner we
looked at the Web. I was once told that the Web meant the death of contact
centers. Obviously, the people who predicted this were dead wrong. Perhaps
speech technology will lower the barriers to entry in starting a company as
it will require fewer agents to be hired. As such, perhaps we will see a
steady stream of newly created companies hiring those out-of-work contact
center agents. Time will tell.
Enough with the predictions. As you'll see in the next pages, I am really
excited about the vendors with which I met to prepare this column. I was
able to sit down with various speech vendors, both at shows and in our TMC
offices, and I'd like to present you with a summary of who is doing what.
Hosted Speech
MCI (www.mci.com) is one of several hosted
speech providers that has in the past been involved in IVR technology and is
now branching out. The company is doing well selling hosted solutions to the
government as well as financial services and insurance companies. It is
currently working on a robust service creation environment that will allow
users to write their own apps and have MCI host them. We can expect to see
biometric (voice authentication) support in the future, as well.
Another option to consider when thinking of hosted solutions is the Voice
Harbor service from CPT International (www.cptii.com).
The company focuses on VoiceXML, speech and multimodal applications and
works with VoiceGenie, ScanSoft and others.
Speech For Developers
There is a great deal of activity in creating tools for speech developers.
LumenVox (www.lumenvox.com) recently
announced version 5.0 of its Speech Recognition Engine, LV Speech Tuner, and
Speech Driven Information System. There is now support for W3C Speech
Recognition Grammar Specification (SRGS) and Semantic Interpretation for
Speech Recognition (SISR). Developers will also be able to have custom
grammars available on-the-fly and modify the engine and grammar performance
at run time.
About a decade ago there were app gen (application generator) wars in the
CTI, or computer-telephony integration, market. Companies like Parity
Software produced environments that competed with the likes of Artisoft (www.artisoft.com),
Envox (www.envox.com) and Pronexu (www.pronexus.com).
Parity was purchased by the communications group of Intel (www.intel.com)
that used to be a standalone company called Dialogic. Since the acquisition,
I haven't heard much about Parity's products. Artisoft abandoned its
development environment and instead bet the farm on an IP-PBX that has done
well over the years. In fact, at press time, Artisoft announced that it is
purchasing Vertical Networks. Pronexus and Envox are still alive and kicking
and focusing on app gens that do speech.
Envox tells me that business is doing well, as companies are looking to
replace proprietary IVR systems with their open solutions, which consist of
products such as Envox 6 Studio, Communications Server, Console and Domain
Server. Their solutions allow users to access databases, e-mail and fax
services, and their development GUI produces Java server pages that
dynamically generate VoiceXML 2.0 code. They can scale to thousands of ports
and use a fault-tolerate run-time engine.
Speech applications are complicated. Let's get that fact straight right from
the start. Most developers aren't experts at developing speech applications
and, as a result, it is not easy to develop complicated solutions. This is
why a company called Tuvox (www.tuvox.com)
has developed what it calls the Perfect Router with a suite of capabilities
that will allow users to manage the entire lifecycle of speech applications.
The system allows companies to develop apps with a GUI; it then helps with
testing and deployment. Once the application is created, users can check for
the hot spots or problem areas and modify the dialogs that will guide
customers so they can better interact with the system.
Speech Meets VoIP
You have probably read my views on the speech and VoIP markets. This is
where 90 percent of the growth in communications (if not all information
technology) is occurring. Vocomo (www.vocomosoft.com)
has combined speech and VoIP by offering a software-based VoIP IVR system.
There are no traditional Dialogic boards; instead, the company used
something called host media processing or HMP, which allows the CPU to
handle what used to require a dedicated DSP resource board. You would
purchase SIP-based 800 numbers from your VoIP service provider and let the
VocomoVoice Response for VoIP system do the rest. The capacity is
impressive. You can have 120 ports per dual CPU server, and the cost is
about $700 per port.
Voice As Brand
Do you have a brand? Yes, you do. Do you need to reflect that brand to your
customers when you interact with them? According to GM Voices (www.gmvoices.com),
you need to think about your voice brand on the phone and on the Web.
Whether you have an IVR system or audio on your Web site, the company argues
that you need to make use of acting talent to perfect your voice brand. Your
brand is the most critical link to your customer and, as a result, you need
to hire professionals when you are working on developing your brand. Visit
the company's Web site for more information.
See You At Speech-World
When you read this, I hope you are making plans to come to TMC's first-ever
Speech-World Conference (www.speech-world.com),
November 30th through December 2nd in Dallas, TX. This is the world's only
speech event that is collocated with an IP contact center conference. We've
got some big name companies backing this event (Cisco Systems, SER,
Inter-Tel and Nuasis to name a few). Speakers will include representatives
from companies such as Genesys, Unisys, LumenVox, Intervoice, Pronexus,
Vocent, Brooktrout, VoiceGenie and Edify.
I am very excited about what speech technologies can do for the call center.
Think back on what ACDs, voice mail and predictive dialing have done in the
past, then consider if you can afford to miss out on this new technology in
your organization. The buzz about this new event is already mounting. We are
looking forward to seeing you there.
Sincerely,
Rich Tehrani
Group Publisher, Group Editor-in-Chief
[email protected]
[
Return To The October 2004 Table Of Contents ] |